ADS1230: ADS1230 and ADS1231

Hi Pranav,

The ADS1230 and the ADS1231 are very similar devices as to device operation.  The ADS1231 includes the capability for a low-side switch to break the excitation current to conserve power.  The ADS1231 has a fixed gain of 128, whereas the ADS1230 allows for a choice of gains 64 or 128.  The ADS1230 has a self-offset calibration whereas the ADS1231 does not.  The ADS1231 has a resolution of 24-bits and the ADS1230 has a resolution of 20-bits. 

Resolution determines the smallest value the ADC can resolve to and does not necessarily guarantee that you will ever achieve this value in a circuit.  There are several reasons and they relate to noise.  One is quantization noise where there is some uncertainty converting a specific analog input value to a digital code.  Noise in the reference voltage adds to the uncertainty.  Another noise source relates to the input signal path and any EMI/RFI noise introduced and then gained up.  Additional noise can be introduced in the PCB layout (and is even worse in a prototype circuit such as a breadboard) from interfering digital noise adjacent to the low analog signal. 

The best possible noise performance that the ADC can achieve is given in the noise tables within the device datasheets.  The lowest noise is achieved with 10sps.  Looking at the noise tables in the datasheet you will find that the noise with a 5V supply/reference is slightly better with the ADS1230 and with 3V supply it is slightly better with the ADS1231.  With load cell applications you are concerned about stable counts and this corresponds with the peak-to-peak noise in the tables.  Scale resolution is a much smaller portion as the output of the load cell is only a small portion of the full-scale range of the ADC.  For this reason you will get the best resolution using the 5V supply as compared to the 3V supply.

Adding factors such as EMI/RFI or power line-cycle noise can lower the resolution even further.  For this reason prototyping solutions often yield poor performance.  So the number of bits that the converter can resolve to is not the best indicator for scale resolution.  You need to determine the noise of the system and the full-scale range relative to the full-scale output of the load cell to determine the scale resolution.

Delta-sigma ADCs offer higher precision by oversampling the input then using a modulator(s) output to push the quantization noise into a higher frequency domain followed by low-pass digital filter.  The low-pass filter cuts out the high frequency noise for a lower noise output.  For this reason you will see a lower noise in the tables for 10sps as opposed to 80sps.  There is an added advantage for the 10sps as the sinc4 digital filter response adds notches at the data rate and multiples of the data rate to notch out power line-cycle noise.

Additional averaging (such as a moving average) can help reduce Gaussian noise, but there is a practical limit of averaging as at some point drift will reduce the benefit.

Perhaps the following document will be of some benefit.  The discussion is with respect to the ADS1232 but the concepts are similar for any load cell or bridge design.

8321.Weigh Scale_Design.doc

Best regards,

Bob B